Face Generation¶

In this project, you'll use generative adversarial networks to generate new images of faces.

Get the Data¶

You'll be using two datasets in this project:

MNIST
CelebA

Since the celebA dataset is complex and you're doing GANs in a project for the first time, we want you to test your neural network on MNIST before CelebA. Running the GANs on MNIST will allow you to see how well your model trains sooner.

If you're using FloydHub, set data_dir to "/input" and use the FloydHub data ID "R5KrjnANiKVhLWAkpXhNBe".

data_dir = './data'

# FloydHub - Use with data ID "R5KrjnANiKVhLWAkpXhNBe"
#data_dir = '/input'


"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import helper

helper.download_extract('mnist', data_dir)
helper.download_extract('celeba', data_dir)

Found mnist Data
Found celeba Data

Explore the Data¶

MNIST¶

As you're aware, the MNIST dataset contains images of handwritten digits. You can view the first number of examples by changing show_n_images.

show_n_images = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
%matplotlib inline
import os
from glob import glob
from matplotlib import pyplot

mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'mnist/*.jpg'))[:show_n_images], 28, 28, 'L')
pyplot.imshow(helper.images_square_grid(mnist_images, 'L'), cmap='gray')

<matplotlib.image.AxesImage at 0x2378458abe0>

CelebA¶

The CelebFaces Attributes Dataset (CelebA) dataset contains over 200,000 celebrity images with annotations. Since you're going to be generating faces, you won't need the annotations. You can view the first number of examples by changing show_n_images.

show_n_images = 25

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
mnist_images = helper.get_batch(glob(os.path.join(data_dir, 'img_align_celeba/*.jpg'))[:show_n_images], 28, 28, 'RGB')
pyplot.imshow(helper.images_square_grid(mnist_images, 'RGB'))

<matplotlib.image.AxesImage at 0x23784638ef0>

Preprocess the Data¶

Since the project's main focus is on building the GANs, we'll preprocess the data for you. The values of the MNIST and CelebA dataset will be in the range of -0.5 to 0.5 of 28x28 dimensional images. The CelebA images will be cropped to remove parts of the image that don't include a face, then resized down to 28x28.

The MNIST images are black and white images with a single color channel while the CelebA images have 3 color channels (RGB color channel).

Build the Neural Network¶

You'll build the components necessary to build a GANs by implementing the following functions below:

model_inputs
discriminator
generator
model_loss
model_opt
train

Check the Version of TensorFlow and Access to GPU¶

This will check to make sure you have the correct version of TensorFlow and access to a GPU

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
from distutils.version import LooseVersion
import warnings
import tensorflow as tf

# Check TensorFlow Version
assert LooseVersion(tf.__version__) >= LooseVersion('1.0'), 'Please use TensorFlow version 1.0 or newer.  You are using {}'.format(tf.__version__)
print('TensorFlow Version: {}'.format(tf.__version__))

# Check for a GPU
if not tf.test.gpu_device_name():
    warnings.warn('No GPU found. Please use a GPU to train your neural network.')
else:
    print('Default GPU Device: {}'.format(tf.test.gpu_device_name()))

D:\Development\Anaconda\envs\dog-windows-gpu\lib\site-packages\h5py\__init__.py:34: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.
  from ._conv import register_converters as _register_converters

TensorFlow Version: 1.7.0
Default GPU Device: /device:GPU:0

Input¶

Implement the model_inputs function to create TF Placeholders for the Neural Network. It should create the following placeholders:

Real input images placeholder with rank 4 using image_width, image_height, and image_channels.
Z input placeholder with rank 2 using z_dim.
Learning rate placeholder with rank 0.

Return the placeholders in the following the tuple (tensor of real input images, tensor of z data)

import problem_unittests as tests

def model_inputs(image_width, image_height, image_channels, z_dim):
    """
    Create the model inputs
    :param image_width: The input image width
    :param image_height: The input image height
    :param image_channels: The number of image channels
    :param z_dim: The dimension of Z
    :return: Tuple of (tensor of real input images, tensor of z data, learning rate)
    """
    # TODO: Implement Function
    inputs_real = tf.placeholder(tf.float32, (None, image_width, image_height, image_channels), name='input_real')
    inputs_z = tf.placeholder(tf.float32, (None, z_dim), name='input_z')
    learning_rate = tf.placeholder(tf.float32, name='learning_rate')

    return inputs_real, inputs_z, learning_rate


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_inputs(model_inputs)

Tests Passed

Discriminator¶

Implement discriminator to create a discriminator neural network that discriminates on images. This function should be able to reuse the variables in the neural network. Use tf.variable_scope with a scope name of "discriminator" to allow the variables to be reused. The function should return a tuple of (tensor output of the discriminator, tensor logits of the discriminator).

def discriminator(images, reuse=False):
    """
    Create the discriminator network
    :param images: Tensor of input image(s)
    :param reuse: Boolean if the weights should be reused
    :return: Tuple of (tensor output of the discriminator, tensor logits of the discriminator)
    """
    # TODO: Implement Function
    alpha = 0 # RELU
    dropout_rate = 0.1    
    with tf.variable_scope('discriminator', reuse=reuse):
        # Input layer is 28x28x3
        x1 = tf.layers.conv2d(images, 32, 5, strides = 2, padding='same')
        relu1 = tf.maximum(alpha*x1, x1)
        #14x14x32
        
        x2 = tf.layers.conv2d(relu1, 64, 5, strides = 2, padding='same')
        bn2 = tf.layers.batch_normalization(x2, training=True)
        relu2 = tf.maximum(alpha*bn2, bn2)
        #7x7x64

        x3 = tf.layers.conv2d(relu2, 128, 5, strides = 1, padding='same')
        bn3 = tf.layers.batch_normalization(x3, training=True)
        relu3 = tf.maximum(alpha*bn3, bn3)
        #7x7x128
                
        flat = tf.reshape(relu3, (-1, 7*7*128))
        flat_with_dropout = tf.layers.dropout(flat,dropout_rate)
        
        logits = tf.layers.dense(flat_with_dropout,1)
        out = tf.sigmoid(logits)
        
        return out, logits


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_discriminator(discriminator, tf)

Tests Passed

Generator¶

Implement generator to generate an image using z. This function should be able to reuse the variables in the neural network. Use tf.variable_scope with a scope name of "generator" to allow the variables to be reused. The function should return the generated 28 x 28 x out_channel_dim images.

def generator(z, out_channel_dim, is_train=True):
    """
    Create the generator network
    :param z: Input z
    :param out_channel_dim: The number of channels in the output image
    :param is_train: Boolean if generator is being used for training
    :return: The tensor output of the generator
    """
    # TODO: Implement Function
    alpha = 0.2 #Leaky RELU factor
    with tf.variable_scope('generator', reuse=not is_train):
        x1 = tf.layers.dense(z, z.shape[1])
        
        #Reshape it to start the convolutional stack
        x1 = tf.reshape(x1, (-1, 7,7,128))
        x1 = tf.layers.batch_normalization(x1, training=is_train)
        x1 = tf.maximum(alpha*x1, x1 )#Leaky RELU
        #128
        
        x2 = tf.layers.conv2d_transpose(x1, 64, 5, strides=1, padding='same')
        x2 = tf.layers.batch_normalization(x2, training=is_train)
        x2 = tf.maximum(alpha*x2, x2)#Leaky RELU
        #64
        
        x3 = tf.layers.conv2d_transpose(x2, 32, 5, strides=2, padding='same')
        x3 = tf.layers.batch_normalization(x3, training=is_train)
        x3 = tf.maximum(alpha*x3, x3)#Leaky RELU
        #32

        #Output layer , 28x28x3
        logits = tf.layers.conv2d_transpose(x3, out_channel_dim, 5, strides=2, padding='same')
        
        out = tf.tanh(logits)
        
        return out


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_generator(generator, tf)

Tests Passed

Loss¶

Implement model_loss to build the GANs for training and calculate the loss. The function should return a tuple of (discriminator loss, generator loss). Use the following functions you implemented:

discriminator(images, reuse=False)
generator(z, out_channel_dim, is_train=True)

def model_loss(input_real, input_z, out_channel_dim):
    """
    Get the loss for the discriminator and generator
    :param input_real: Images from the real dataset
    :param input_z: Z input
    :param out_channel_dim: The number of channels in the output image
    :return: A tuple of (discriminator loss, generator loss)
    """
    # TODO: Implement Function
    g_model = generator(input_z, out_channel_dim)
    d_model_real, d_logits_real = discriminator(input_real)
    d_model_fake, d_logits_fake = discriminator(g_model, reuse=True)
    #Label Smoothing factor - To prevent discriminator from being too strong 
    smooth = 0.1
    
    d_loss_real = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_real, labels=tf.ones_like(d_model_real)*(1 - smooth)))
    d_loss_fake = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake, labels=tf.zeros_like(d_model_fake)))
    g_loss = tf.reduce_mean(
        tf.nn.sigmoid_cross_entropy_with_logits(logits=d_logits_fake, labels=tf.ones_like(d_model_fake)))

    d_loss = d_loss_real + d_loss_fake

    return d_loss, g_loss

"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_loss(model_loss)

Tests Passed

Optimization¶

Implement model_opt to create the optimization operations for the GANs. Use tf.trainable_variables to get all the trainable variables. Filter the variables with names that are in the discriminator and generator scope names. The function should return a tuple of (discriminator training operation, generator training operation).

def model_opt(d_loss, g_loss, learning_rate, beta1):
    """
    Get optimization operations
    :param d_loss: Discriminator loss Tensor
    :param g_loss: Generator loss Tensor
    :param learning_rate: Learning Rate Placeholder
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :return: A tuple of (discriminator training operation, generator training operation)
    """
    # TODO: Implement Function
    t_vars = tf.trainable_variables()
    d_vars = [var for var in t_vars if var.name.startswith('discriminator')]
    g_vars = [var for var in t_vars if var.name.startswith('generator')]

    # Optimize
    with tf.control_dependencies(tf.get_collection(tf.GraphKeys.UPDATE_OPS)):
        d_train_opt = tf.train.AdamOptimizer(learning_rate, beta1=beta1).minimize(d_loss, var_list=d_vars)
        g_train_opt = tf.train.AdamOptimizer(learning_rate, beta1=beta1).minimize(g_loss, var_list=g_vars)

    return d_train_opt, g_train_opt    


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
tests.test_model_opt(model_opt, tf)

Tests Passed

Neural Network Training¶

Show Output¶

Use this function to show the current output of the generator during training. It will help you determine how well the GANs is training.

"""
DON'T MODIFY ANYTHING IN THIS CELL
"""
import numpy as np

def show_generator_output(sess, n_images, input_z, out_channel_dim, image_mode):
    """
    Show example output for the generator
    :param sess: TensorFlow session
    :param n_images: Number of Images to display
    :param input_z: Input Z Tensor
    :param out_channel_dim: The number of channels in the output image
    :param image_mode: The mode to use for images ("RGB" or "L")
    """
    cmap = None if image_mode == 'RGB' else 'gray'
    z_dim = input_z.get_shape().as_list()[-1]
    example_z = np.random.uniform(-1, 1, size=[n_images, z_dim])

    samples = sess.run(
        generator(input_z, out_channel_dim, False),
        feed_dict={input_z: example_z})

    images_grid = helper.images_square_grid(samples, image_mode)
    pyplot.imshow(images_grid, cmap=cmap)
    pyplot.show()

Train¶

Implement train to build and train the GANs. Use the following functions you implemented:

model_inputs(image_width, image_height, image_channels, z_dim)
model_loss(input_real, input_z, out_channel_dim)
model_opt(d_loss, g_loss, learning_rate, beta1)

Use the show_generator_output to show generator output while you train. Running show_generator_output for every batch will drastically increase training time and increase the size of the notebook. It's recommended to print the generator output every 100 batches.

def scale(x, feature_range=(-1, 1)):
    # scale to (0, 1)
    x = ((x - x.min())/(255 - x.min()))
    
    # scale to feature_range
    min, max = feature_range
    x = x * (max - min) + min
    return x

def train(epoch_count, batch_size, z_dim, learning_rate, beta1, get_batches, data_shape, data_image_mode):
    """
    Train the GAN
    :param epoch_count: Number of epochs
    :param batch_size: Batch Size
    :param z_dim: Z dimension
    :param learning_rate: Learning Rate
    :param beta1: The exponential decay rate for the 1st moment in the optimizer
    :param get_batches: Function to get batches
    :param data_shape: Shape of the data
    :param data_image_mode: The image mode to use for images ("RGB" or "L")
    """
    # TODO: Build Model
    
    print_every = 10
    show_every=100
    steps = 0

    inputs_real, inputs_z, learning_rate_var = model_inputs(data_shape[1], data_shape[2], data_shape[3], z_dim)
    d_loss, g_loss = model_loss(inputs_real, inputs_z, data_shape[3])
    d_train_opt, g_train_opt  = model_opt(d_loss, g_loss, learning_rate, beta1)
    
    config = tf.ConfigProto()
    config.gpu_options.allocator_type = 'BFC'
    #config.gpu_options.allow_growth = True
    config.gpu_options.per_process_gpu_memory_fraction = 0.9 # This is to make sure that the GPU does not run out of Memory.

    
    with tf.Session(config=config) as sess:
        sess.run(tf.global_variables_initializer())
        for epoch_i in range(epoch_count):
            for batch_images in get_batches(batch_size):
                #batch_images = scale(batch_images) #Scale images between -1 to 1 to match tanh input from generator. Commented this as it was resulting in bad images
                steps += 1
                # TODO: Train Model
                # Run optimizers                
                batch_z = np.random.normal(0, 1, size=(batch_size, z_dim))
                
                _ = sess.run(d_train_opt, feed_dict={inputs_real: batch_images, inputs_z: batch_z, learning_rate_var: learning_rate})
                _ = sess.run(g_train_opt, feed_dict={inputs_real: batch_images, inputs_z: batch_z, learning_rate_var: learning_rate})
                
                if steps % print_every == 0:
                    train_loss_d = d_loss.eval({inputs_real: batch_images, inputs_z: batch_z, learning_rate_var: learning_rate})
                    train_loss_g = g_loss.eval({inputs_real: batch_images, inputs_z: batch_z, learning_rate_var: learning_rate})
                    print("Epoch {}/{}...".format(epoch_i+1, epoch_count),
                          "Discriminator Loss: {:.4f}...".format(train_loss_d),
                          "Generator Loss: {:.4f}".format(train_loss_g))
                    
                if steps % show_every == 0:                    
                    show_generator_output(sess, 25, inputs_z, data_shape[3], data_image_mode)

MNIST¶

Test your GANs architecture on MNIST. After 2 epochs, the GANs should be able to generate images that look like handwritten digits. Make sure the loss of the generator is lower than the loss of the discriminator or close to 0.

batch_size = 16
z_dim = 6272
learning_rate = 0.0002
beta1 = 0.1 # AS per review, optimal value is between 0.1 and 0.3 for good results


"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 2

mnist_dataset = helper.Dataset('mnist', glob(os.path.join(data_dir, 'mnist/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, mnist_dataset.get_batches,
          mnist_dataset.shape, mnist_dataset.image_mode)

Epoch 1/2... Discriminator Loss: 2.1024... Generator Loss: 0.2198
Epoch 1/2... Discriminator Loss: 1.5175... Generator Loss: 0.4695
Epoch 1/2... Discriminator Loss: 1.2716... Generator Loss: 0.6739
Epoch 1/2... Discriminator Loss: 2.3192... Generator Loss: 0.1880
Epoch 1/2... Discriminator Loss: 2.4522... Generator Loss: 0.1569
Epoch 1/2... Discriminator Loss: 1.5848... Generator Loss: 0.5669
Epoch 1/2... Discriminator Loss: 3.9107... Generator Loss: 0.0365
Epoch 1/2... Discriminator Loss: 1.4119... Generator Loss: 0.6383
Epoch 1/2... Discriminator Loss: 2.6375... Generator Loss: 0.1274
Epoch 1/2... Discriminator Loss: 1.2823... Generator Loss: 0.6269

Epoch 1/2... Discriminator Loss: 2.1078... Generator Loss: 0.2194
Epoch 1/2... Discriminator Loss: 2.5783... Generator Loss: 0.1251
Epoch 1/2... Discriminator Loss: 1.4670... Generator Loss: 0.4260
Epoch 1/2... Discriminator Loss: 0.8208... Generator Loss: 1.2991
Epoch 1/2... Discriminator Loss: 3.0608... Generator Loss: 0.0772
Epoch 1/2... Discriminator Loss: 2.3659... Generator Loss: 0.1568
Epoch 1/2... Discriminator Loss: 1.0533... Generator Loss: 0.9834
Epoch 1/2... Discriminator Loss: 3.3106... Generator Loss: 0.0653
Epoch 1/2... Discriminator Loss: 2.9405... Generator Loss: 0.1534
Epoch 1/2... Discriminator Loss: 1.8566... Generator Loss: 1.1870

Epoch 1/2... Discriminator Loss: 3.5780... Generator Loss: 0.0506
Epoch 1/2... Discriminator Loss: 2.1090... Generator Loss: 0.2336
Epoch 1/2... Discriminator Loss: 1.8535... Generator Loss: 0.3095
Epoch 1/2... Discriminator Loss: 1.2753... Generator Loss: 0.8038
Epoch 1/2... Discriminator Loss: 1.3615... Generator Loss: 0.6025
Epoch 1/2... Discriminator Loss: 1.1932... Generator Loss: 1.0226
Epoch 1/2... Discriminator Loss: 0.6954... Generator Loss: 1.6479
Epoch 1/2... Discriminator Loss: 1.4227... Generator Loss: 1.8983
Epoch 1/2... Discriminator Loss: 0.9711... Generator Loss: 1.2930
Epoch 1/2... Discriminator Loss: 1.0481... Generator Loss: 1.0154

Epoch 1/2... Discriminator Loss: 0.9155... Generator Loss: 1.1848
Epoch 1/2... Discriminator Loss: 1.3866... Generator Loss: 0.6242
Epoch 1/2... Discriminator Loss: 1.2498... Generator Loss: 0.6885
Epoch 1/2... Discriminator Loss: 1.1743... Generator Loss: 0.8942
Epoch 1/2... Discriminator Loss: 1.9527... Generator Loss: 0.3095
Epoch 1/2... Discriminator Loss: 1.2335... Generator Loss: 0.7014
Epoch 1/2... Discriminator Loss: 1.9254... Generator Loss: 0.2940
Epoch 1/2... Discriminator Loss: 1.9367... Generator Loss: 0.3496
Epoch 1/2... Discriminator Loss: 0.7883... Generator Loss: 1.2623
Epoch 1/2... Discriminator Loss: 1.1298... Generator Loss: 0.7123

Epoch 1/2... Discriminator Loss: 2.8441... Generator Loss: 0.1601
Epoch 1/2... Discriminator Loss: 0.9876... Generator Loss: 0.9338
Epoch 1/2... Discriminator Loss: 1.6571... Generator Loss: 0.4198
Epoch 1/2... Discriminator Loss: 1.2659... Generator Loss: 0.6460
Epoch 1/2... Discriminator Loss: 0.8050... Generator Loss: 1.2337
Epoch 1/2... Discriminator Loss: 0.6970... Generator Loss: 1.3307
Epoch 1/2... Discriminator Loss: 1.2447... Generator Loss: 0.6885
Epoch 1/2... Discriminator Loss: 0.6812... Generator Loss: 1.8139
Epoch 1/2... Discriminator Loss: 2.3575... Generator Loss: 0.2290
Epoch 1/2... Discriminator Loss: 1.3631... Generator Loss: 0.6944

Epoch 1/2... Discriminator Loss: 0.9051... Generator Loss: 1.0248
Epoch 1/2... Discriminator Loss: 1.2847... Generator Loss: 0.6854
Epoch 1/2... Discriminator Loss: 1.1889... Generator Loss: 0.7854
Epoch 1/2... Discriminator Loss: 0.7581... Generator Loss: 1.5189
Epoch 1/2... Discriminator Loss: 0.7159... Generator Loss: 1.4230
Epoch 1/2... Discriminator Loss: 0.8468... Generator Loss: 1.0373
Epoch 1/2... Discriminator Loss: 0.7659... Generator Loss: 1.3036
Epoch 1/2... Discriminator Loss: 0.9524... Generator Loss: 0.9550
Epoch 1/2... Discriminator Loss: 0.7081... Generator Loss: 1.6391
Epoch 1/2... Discriminator Loss: 0.6715... Generator Loss: 2.1343

Epoch 1/2... Discriminator Loss: 0.6913... Generator Loss: 1.4339
Epoch 1/2... Discriminator Loss: 0.7207... Generator Loss: 1.6424
Epoch 1/2... Discriminator Loss: 0.7698... Generator Loss: 1.5515
Epoch 1/2... Discriminator Loss: 1.1077... Generator Loss: 0.7269
Epoch 1/2... Discriminator Loss: 1.0613... Generator Loss: 0.8271
Epoch 1/2... Discriminator Loss: 1.4280... Generator Loss: 0.6735
Epoch 1/2... Discriminator Loss: 1.0662... Generator Loss: 0.8558
Epoch 1/2... Discriminator Loss: 1.6137... Generator Loss: 0.5738
Epoch 1/2... Discriminator Loss: 1.9849... Generator Loss: 0.2915
Epoch 1/2... Discriminator Loss: 0.7328... Generator Loss: 1.3233

Epoch 1/2... Discriminator Loss: 0.8120... Generator Loss: 1.2011
Epoch 1/2... Discriminator Loss: 0.9322... Generator Loss: 1.1330
Epoch 1/2... Discriminator Loss: 0.6375... Generator Loss: 1.5504
Epoch 1/2... Discriminator Loss: 0.8198... Generator Loss: 1.1615
Epoch 1/2... Discriminator Loss: 0.7816... Generator Loss: 3.4389
Epoch 1/2... Discriminator Loss: 0.8266... Generator Loss: 1.0727
Epoch 1/2... Discriminator Loss: 0.5785... Generator Loss: 2.1341
Epoch 1/2... Discriminator Loss: 0.7105... Generator Loss: 1.4014
Epoch 1/2... Discriminator Loss: 0.8464... Generator Loss: 1.2733
Epoch 1/2... Discriminator Loss: 0.5616... Generator Loss: 1.8185

Epoch 1/2... Discriminator Loss: 0.5590... Generator Loss: 2.0826
Epoch 1/2... Discriminator Loss: 0.6964... Generator Loss: 3.3680
Epoch 1/2... Discriminator Loss: 0.5264... Generator Loss: 2.1821
Epoch 1/2... Discriminator Loss: 4.3745... Generator Loss: 0.0348
Epoch 1/2... Discriminator Loss: 0.9967... Generator Loss: 0.8799
Epoch 1/2... Discriminator Loss: 0.6648... Generator Loss: 3.6419
Epoch 1/2... Discriminator Loss: 0.5546... Generator Loss: 2.5779
Epoch 1/2... Discriminator Loss: 0.5668... Generator Loss: 2.0540
Epoch 1/2... Discriminator Loss: 0.4649... Generator Loss: 2.3491
Epoch 1/2... Discriminator Loss: 0.5162... Generator Loss: 2.5768

Epoch 1/2... Discriminator Loss: 0.5129... Generator Loss: 2.5867
Epoch 1/2... Discriminator Loss: 1.2860... Generator Loss: 0.8641
Epoch 1/2... Discriminator Loss: 0.4109... Generator Loss: 3.1209
Epoch 1/2... Discriminator Loss: 0.3622... Generator Loss: 4.6959
Epoch 1/2... Discriminator Loss: 0.5245... Generator Loss: 2.3533
Epoch 1/2... Discriminator Loss: 0.4706... Generator Loss: 2.8002
Epoch 1/2... Discriminator Loss: 0.3730... Generator Loss: 3.8006
Epoch 1/2... Discriminator Loss: 0.5126... Generator Loss: 2.3143
Epoch 1/2... Discriminator Loss: 0.4874... Generator Loss: 2.6855
Epoch 1/2... Discriminator Loss: 0.3723... Generator Loss: 4.0798

CelebA¶

Run your GANs on CelebA. It will take around 20 minutes on the average GPU to run one epoch. You can run the whole epoch or stop when it starts to generate realistic faces.

batch_size = 16
z_dim = 6272
learning_rate = 0.0002
beta1 = 0.1



"""
DON'T MODIFY ANYTHING IN THIS CELL THAT IS BELOW THIS LINE
"""
epochs = 1

celeba_dataset = helper.Dataset('celeba', glob(os.path.join(data_dir, 'img_align_celeba/*.jpg')))
with tf.Graph().as_default():
    train(epochs, batch_size, z_dim, learning_rate, beta1, celeba_dataset.get_batches,
          celeba_dataset.shape, celeba_dataset.image_mode)

Epoch 1/1... Discriminator Loss: 4.2386... Generator Loss: 0.0294
Epoch 1/1... Discriminator Loss: 3.6241... Generator Loss: 0.0500
Epoch 1/1... Discriminator Loss: 2.9555... Generator Loss: 0.1116
Epoch 1/1... Discriminator Loss: 2.4076... Generator Loss: 0.1652
Epoch 1/1... Discriminator Loss: 2.1988... Generator Loss: 0.1978
Epoch 1/1... Discriminator Loss: 3.0723... Generator Loss: 0.0873
Epoch 1/1... Discriminator Loss: 2.6618... Generator Loss: 0.1236
Epoch 1/1... Discriminator Loss: 2.3067... Generator Loss: 0.1939
Epoch 1/1... Discriminator Loss: 2.9930... Generator Loss: 0.0961
Epoch 1/1... Discriminator Loss: 2.1163... Generator Loss: 0.2155

Epoch 1/1... Discriminator Loss: 1.8780... Generator Loss: 0.2703
Epoch 1/1... Discriminator Loss: 2.7281... Generator Loss: 0.1654
Epoch 1/1... Discriminator Loss: 2.5709... Generator Loss: 0.1498
Epoch 1/1... Discriminator Loss: 1.6365... Generator Loss: 0.3646
Epoch 1/1... Discriminator Loss: 2.6826... Generator Loss: 0.1475
Epoch 1/1... Discriminator Loss: 2.2015... Generator Loss: 0.1961
Epoch 1/1... Discriminator Loss: 1.3988... Generator Loss: 0.4967
Epoch 1/1... Discriminator Loss: 2.6671... Generator Loss: 0.1163
Epoch 1/1... Discriminator Loss: 2.1896... Generator Loss: 0.2135
Epoch 1/1... Discriminator Loss: 3.5493... Generator Loss: 0.0636

Epoch 1/1... Discriminator Loss: 4.1255... Generator Loss: 0.0400
Epoch 1/1... Discriminator Loss: 2.1266... Generator Loss: 0.2642
Epoch 1/1... Discriminator Loss: 1.4698... Generator Loss: 1.3356
Epoch 1/1... Discriminator Loss: 4.4252... Generator Loss: 0.0408
Epoch 1/1... Discriminator Loss: 3.1510... Generator Loss: 0.1432
Epoch 1/1... Discriminator Loss: 2.1027... Generator Loss: 0.3462
Epoch 1/1... Discriminator Loss: 2.5319... Generator Loss: 0.2549
Epoch 1/1... Discriminator Loss: 0.6719... Generator Loss: 2.0376
Epoch 1/1... Discriminator Loss: 1.1478... Generator Loss: 0.8804
Epoch 1/1... Discriminator Loss: 0.9670... Generator Loss: 1.4345

Epoch 1/1... Discriminator Loss: 0.8631... Generator Loss: 2.2558
Epoch 1/1... Discriminator Loss: 1.3231... Generator Loss: 0.9835
Epoch 1/1... Discriminator Loss: 0.7391... Generator Loss: 1.9310
Epoch 1/1... Discriminator Loss: 0.8743... Generator Loss: 1.4682
Epoch 1/1... Discriminator Loss: 0.7496... Generator Loss: 1.5691
Epoch 1/1... Discriminator Loss: 0.6614... Generator Loss: 1.5597
Epoch 1/1... Discriminator Loss: 0.8479... Generator Loss: 1.2571
Epoch 1/1... Discriminator Loss: 1.2939... Generator Loss: 0.7458
Epoch 1/1... Discriminator Loss: 1.9484... Generator Loss: 0.2789
Epoch 1/1... Discriminator Loss: 0.8126... Generator Loss: 1.1756

Epoch 1/1... Discriminator Loss: 0.6280... Generator Loss: 1.9838
Epoch 1/1... Discriminator Loss: 0.3624... Generator Loss: 4.6729
Epoch 1/1... Discriminator Loss: 2.1630... Generator Loss: 0.2427
Epoch 1/1... Discriminator Loss: 2.3202... Generator Loss: 0.1769
Epoch 1/1... Discriminator Loss: 1.2705... Generator Loss: 0.6447
Epoch 1/1... Discriminator Loss: 2.6640... Generator Loss: 0.1390
Epoch 1/1... Discriminator Loss: 1.1361... Generator Loss: 0.7013
Epoch 1/1... Discriminator Loss: 1.2263... Generator Loss: 0.7084
Epoch 1/1... Discriminator Loss: 1.0901... Generator Loss: 0.9530
Epoch 1/1... Discriminator Loss: 1.4204... Generator Loss: 0.6985

Epoch 1/1... Discriminator Loss: 0.7984... Generator Loss: 1.5864
Epoch 1/1... Discriminator Loss: 2.0578... Generator Loss: 0.4309
Epoch 1/1... Discriminator Loss: 0.8178... Generator Loss: 3.3107
Epoch 1/1... Discriminator Loss: 0.7424... Generator Loss: 1.6308
Epoch 1/1... Discriminator Loss: 1.5358... Generator Loss: 0.5458
Epoch 1/1... Discriminator Loss: 1.1318... Generator Loss: 0.8018
Epoch 1/1... Discriminator Loss: 1.6372... Generator Loss: 0.4283
Epoch 1/1... Discriminator Loss: 4.7121... Generator Loss: 0.0190
Epoch 1/1... Discriminator Loss: 1.0529... Generator Loss: 0.8219
Epoch 1/1... Discriminator Loss: 1.4198... Generator Loss: 4.3442

Epoch 1/1... Discriminator Loss: 1.2396... Generator Loss: 0.5767
Epoch 1/1... Discriminator Loss: 2.4004... Generator Loss: 1.9683
Epoch 1/1... Discriminator Loss: 1.2722... Generator Loss: 0.5879
Epoch 1/1... Discriminator Loss: 3.4495... Generator Loss: 0.0597
Epoch 1/1... Discriminator Loss: 1.0648... Generator Loss: 1.0426
Epoch 1/1... Discriminator Loss: 0.7563... Generator Loss: 1.3561
Epoch 1/1... Discriminator Loss: 1.2126... Generator Loss: 0.8008
Epoch 1/1... Discriminator Loss: 1.9417... Generator Loss: 0.2668
Epoch 1/1... Discriminator Loss: 0.6826... Generator Loss: 2.2821
Epoch 1/1... Discriminator Loss: 1.0872... Generator Loss: 0.9321

Epoch 1/1... Discriminator Loss: 1.1296... Generator Loss: 0.7257
Epoch 1/1... Discriminator Loss: 1.8312... Generator Loss: 0.3077
Epoch 1/1... Discriminator Loss: 1.1144... Generator Loss: 0.7166
Epoch 1/1... Discriminator Loss: 1.1389... Generator Loss: 0.7046
Epoch 1/1... Discriminator Loss: 1.5551... Generator Loss: 0.5141
Epoch 1/1... Discriminator Loss: 1.7922... Generator Loss: 0.3376
Epoch 1/1... Discriminator Loss: 0.7563... Generator Loss: 1.6917
Epoch 1/1... Discriminator Loss: 1.4534... Generator Loss: 0.5153
Epoch 1/1... Discriminator Loss: 0.7134... Generator Loss: 1.5111
Epoch 1/1... Discriminator Loss: 1.7121... Generator Loss: 0.4753

Epoch 1/1... Discriminator Loss: 1.4573... Generator Loss: 0.6188
Epoch 1/1... Discriminator Loss: 2.1129... Generator Loss: 0.2714
Epoch 1/1... Discriminator Loss: 1.7772... Generator Loss: 0.3359
Epoch 1/1... Discriminator Loss: 1.4206... Generator Loss: 0.7428
Epoch 1/1... Discriminator Loss: 1.1269... Generator Loss: 0.8930
Epoch 1/1... Discriminator Loss: 3.5975... Generator Loss: 0.0540
Epoch 1/1... Discriminator Loss: 1.0247... Generator Loss: 0.9302
Epoch 1/1... Discriminator Loss: 2.1557... Generator Loss: 0.2520
Epoch 1/1... Discriminator Loss: 1.1835... Generator Loss: 0.7247
Epoch 1/1... Discriminator Loss: 1.5060... Generator Loss: 0.4695

Epoch 1/1... Discriminator Loss: 1.3936... Generator Loss: 0.5052
Epoch 1/1... Discriminator Loss: 1.5765... Generator Loss: 0.5116
Epoch 1/1... Discriminator Loss: 1.1242... Generator Loss: 0.9118
Epoch 1/1... Discriminator Loss: 0.8914... Generator Loss: 1.1948
Epoch 1/1... Discriminator Loss: 0.8144... Generator Loss: 1.1784
Epoch 1/1... Discriminator Loss: 1.2632... Generator Loss: 0.6164
Epoch 1/1... Discriminator Loss: 2.1322... Generator Loss: 0.2360
Epoch 1/1... Discriminator Loss: 2.8307... Generator Loss: 0.1521
Epoch 1/1... Discriminator Loss: 1.6109... Generator Loss: 0.4160
Epoch 1/1... Discriminator Loss: 2.0618... Generator Loss: 0.3940

Submitting This Project¶

When submitting this project, make sure to run all the cells before saving the notebook. Save the notebook file as "dlnd_face_generation.ipynb" and save it as a HTML file under "File" -> "Download as". Include the "helper.py" and "problem_unittests.py" files in your submission.